As I told, major regression for us. So that means that was our biggest issue so far :D
Cheers From: softimage-boun...@listproc.autodesk.com [mailto:softimage-boun...@listproc.autodesk.com] On Behalf Of Raffaele Fragapane Sent: Tuesday, May 07, 2013 10:25 AM To: softimage@listproc.autodesk.com Subject: Re: 2014 in production anyone? I'm not 100% sure I would call something that can be fixed with a script swap, or a 20 lines script bound to the hotkey, a MAJOR regression ;) I know you got super annoyed at that though. 2012 was actually quite a good release for us, but that's after QFEs, 2013 seems to be doing ok too, but we use it more "experimentally" and for batched jobs at the end of the pipe, not user-facing. A lot of the things fixed from production reports though I saw cropping up in 2014, some of them are very, very significant, so that's encouraging. Currently waiting on my license to come in to play at home more with it, at work and home both I'm on 2012 and I don't agree that it's a bad release, quite the opposite in my experience, and we milked it like a pregnant cow on two shows, but as is often the case depending on how and where it's used, other than some disastrous releases, individual mileage varies greatly. I feel sorry for Matt every time I read his posts :p I'm glad I don't have to live with that kind of legacy to support. On Tue, May 7, 2013 at 6:12 PM, Szabolcs Matefy <szabol...@crytek.com> wrote: We switched from 2013 SP1 to 2014. Major regression for us is that when you hide polygons, and select some polygons then unhide the hidden polygons, your selection will be lost. There are several bugs still present, however overall feeling is better. We don't rely on RT shading, but HQV is way better than the previous versions, and there some important bugfixes too. So I suggest you to check your current assets in 2014, work for a week, and if you find no issues, migrate. Cheers From: softimage-boun...@listproc.autodesk.com [mailto:softimage-boun...@listproc.autodesk.com] On Behalf Of Jens Lindgren Sent: Tuesday, May 07, 2013 10:02 AM To: softimage@listproc.autodesk.com Subject: Re: 2014 in production anyone? I had to switch to 2014 for the current project due to a bug in 2013 that crashed Softimage when importing reference models with Ice Trees on it. In 2014 I can import these models as references without problems. One other bonus is the premultiply on LUTs options. Being able to set it to "No" is a really nice addition and I don't have any artifacts with the LUTs anymore. /Jens On Tue, May 7, 2013 at 7:04 AM, Enrique Caballero <enriquecaball...@gmail.com> wrote: Hmm i thought 2012 and 2013 were decent releases. 13 for some reason really slowed down importing of referenced models that had ICE strands in them but 2011 was a very troublesome release, the broken shader issue was painful On Tue, May 7, 2013 at 11:59 AM, Matt Lind <ml...@carbinestudios.com> wrote: Let me follow up and say 2014 is probably a safer bet to use in production than 2013, 2012 or 2011. I would definitely stay away from 2012 as there are some hidden gremlins that won't show up until the data is migrated forward into a newer version. Matt From: softimage-boun...@listproc.autodesk.com [mailto:softimage-boun...@listproc.autodesk.com] On Behalf Of Sylvain Lebeau Sent: Monday, May 06, 2013 8:51 PM To: softimage@listproc.autodesk.com Cc: 3...@shedmtl.com Subject: Re: 2014 in production anyone? My god Matt thanks so much for this bible!!! Altought we dont rely on rt shaders, your manifesto is such a good pointer man.... Much appreciated Thanks a big bunch! Sly Sylvain Lebeau // SHED V-P/Visual effects supervisor 1410, RUE STANLEY, 11E ÉTAGE MONTRÉAL (QUÉBEC) H3A 1P8 T 514 849-1555 <tel:514%20849-1555> F 514 849-5025 <tel:514%20849-5025> WWW.SHEDMTL.COM <http://www.shedmtl.com/> On 2013-05-06, at 22:31, Matt Lind <ml...@carbinestudios.com> wrote: We haven't switched to 2014 yet, but in testing we've encountered a number of regressions in important areas to us. Let me emphasize what is important to us might not be important to you. In our case the problem has been the ability to upgrade and migrate old data into a current release and have it continue to function as expected - e.g. be durable. For example, we use realtime shaders very heavily in our production...in fact every single asset in our production uses them. But there's always some type of bug/regression in each new release with realtime shading preventing an upgrade. We were stuck on 7.5 for years until 2013 SP1 finally resolved the shading issues. However, upon testing everything else we discovered many animation features regressed such as ability to import FCurve data efficiently, rotation order not computed properly on transforms, scene doesn't update/refresh properly, cut n' paste FCurve keys fails, envelopes get really cranky/crash when a deformer is not found, callbacks/events intermittently fail or don't execute, scene layers misbehave, etc... It's been whack-a-mole. Many bugs we've found in 2013 are still present in 2014. Problem is we cannot consistently reproduce some of these bugs as the variables leading to their appearance can be complex much like needing the stars to align a certain way before the issue is revealed. We just don't have a complete picture of all the stars involved in the problem yet to be able to submit a report for further investigation. It's a real stressing point in our production. In regards to 2014, the realtime shading architecture migration from RTS 3 to RTS 4 doesn't work, or there is no migration path. Since RTS 3 is technically deprecated (but still supported) that is a problem for us. If a shader is not installed, Softimage won't open the scene/.emdl. In fact, Softimage will crash in spectacular fashion. If you don't use realtime shaders, then this won't be an issue for you. If I had to summarize the main issues, it seems like many features in recent releases touched on core components and were either inserted in rush fashion or sloppily implemented leading to stupid bugs and regressions. These largely revolve around system updates and notifications on the backend. For the other stuff where it's more data centric, the data gets into scene files and becomes a problem when the data is migrated forward to a newer release and needs to be iterated by an artist. Problem is the data no longer behaves the same. We've had a few models created in 7.5 that when brought into current versions don't react the same way. Either data in the delta is getting interpreted differently, or in some cases, the file just won't load. Envelopes are a victim of this specific problem. I could go on, but the moral of the story is to test thoroughly before upgrading. Of course, that lesson applies to every product. Matt From: softimage-boun...@listproc.autodesk.com [mailto:softimage-boun...@listproc.autodesk.com] On Behalf Of Sylvain Lebeau Sent: Monday, May 06, 2013 4:25 PM To: softimage@listproc.autodesk.com Subject: 2014 in production anyone? Hi everyone.... After flaming ferosiously at the 2014 release, seeing the bug fixes list quite brought me back down to earth. Very extensive list so hat's off to the new devs for this. Must be quite a puzzle. And to me this is much more important then new tools. In the meantime, i would really like to hear about peoples who already switched. What are the culprits you've went into?, backward compatibility?, broken addons you rely upon? Regressions?... every single thing that could make me happy about my renewed subscription would be appreciated. thanks for your time! sly Sylvain Lebeau // SHED V-P/Visual effects supervisor 1410, RUE STANLEY, 11E ÉTAGE MONTRÉAL (QUÉBEC) H3A 1P8 T 514 849-1555 <tel:514%20849-1555> F 514 849-5025 <tel:514%20849-5025> WWW.SHEDMTL.COM <http://www.shedmtl.com/> <http://WWW.SHEDMTL.COM <http://www.shedmtl.com/> > -- Jens Lindgren -------------------------- Lead Technical Director Magoo 3D Studios <http://www.magoo3dstudios.com/> -- Our users will know fear and cower before our software! Ship it! Ship it and let them flee like the dogs they are!
